Error Patterns from Alternative Cost Progress Models

Abstract

Abstract Numerous cost progress models have been proposed in the literature and used in practice. This paper selects five cost progress models, which predict future cost using various combinations of three factors (past cost, cumulative quantity, and production rate), and investigates the forecast accuracy of the alternative models under varying circumstances. The models examined include the random walk model, the traditional learning curve model, and common models designed to incorporate the impact of production rate on cost. The broad objectives are to (1) identify conditions which may affect model accuracy, documenting the manner in which forecast errors for each model depend on those conditions, and (2) suggest which of the five models may be more or less accurate under a given set of conditions. Particular attention is paid to how model accuracy is affected by one specific condition—changes in production rate. Tests of model forecast accuracy are constructed using annual cost and quantity data from a large sample of major aerospace weapon system programs. Findings indicate that a simple random walk model is most accurate under most circumstances, and raises the question of when more sophisticated models may outperform the random walk.